Comp
Comp
a r t i c l e i n f o a b s t r a c t
Article history: The electrical and dielectric properties of Pr0.5-xGdxSr0.5MnO3 (0 x 0.3) manganites prepared by the
Received 29 July 2017 conventional sol-gel method have been studied in a wide range of temperature [80e320 K] and fre-
Received in revised form quency [40e107 Hz] using impedance spectroscopy technique. DC conductivity (sDC) results show that
6 January 2018
for x ¼ 0, 0.1 and 0.2, the samples present a semiconductor behavior in the temperature range of the
Accepted 16 January 2018
study. For x ¼ 0.3, a semiconductor-metal transition is observed at 280 K. sDC shows also a decrease in
the conductivity with the increase of Gd doping. The calculated activation energies are Gd content
dependent. AC conductivity results were analysed using Drude model in the metallic region and by the
Keywords:
Pr0.5Sr0.5MnO3
Jonsher power law in the semiconductor one. The Niquist plots for x ¼ 0.2 and x ¼ 0.3 obeys at Cole-Cole
Gd-doping model. The impedance spectra were fitted by an adequate equivalent circuit involving two contributions
Impedance spectroscopy attributed to grains and grain boundaries associated with an inductance. Such a result reveals a corre-
Electrical properties lation between electrical and magnetic properties of Gd-doped manganites. In addition, grain-boundary
Drude model resistance shows a major contribution at higher temperature. High dielectric constants are obtained in
Dielectric properties particular at low frequencies. Dielectric loss is analyzed through the Giuntini theory with which we have
calculated the energy required for charge carriers to hope over the potential barriers.
© 2018 Elsevier B.V. All rights reserved.
https://doi.org/10.1016/j.jallcom.2018.01.236
0925-8388/© 2018 Elsevier B.V. All rights reserved.
724 A. Ben Jazia Kharrat et al. / Journal of Alloys and Compounds 741 (2018) 723e733
the bulk, grain boundary and electrode contributions. xGdxSr0.5MnO3 (x ¼ 0; 0.1, 0.2 and 0.3) compounds. The obtained
In the present study, we report the effect of gadolinium on the lines are fine and intense and signature of a good crystallization of
temperature and frequency dependence of electrical and dielectric our samples. Using the Rietveld refinement technique with the
properties of polycrystalline Pr0.5-xGdxSr0.5MnO3 (x ¼ 0, 0.1, 0.2 and Fullprof program [17], we have deduced that the parent sample
0.3) compounds prepared by sol-gel route. crystallizes in the tetragonal structure with I4/mcm space group but
the doped ones crystallize in the distorted orthorhombic structure
2. Experimental details with Pbnm space group.
The cell lattice parameters, the unit cell volume V, the average
Polycrystalline samples of Pr0.5-xGdxSr0.5MnO3 (x ¼ 0, 0.1, 0.2 〈MneO〉 bond length, the average angle 〈MneOeMn〉, the average
and 0.3) were prepared using the conventional sol-gel method [16]. ionic radius in A-site 〈rA〉 and the cationic disorder in A-site s2 for
We have used as precursors Pr6O11, Gd2O3, SrCO3 and MnO2 up to all our compounds, deduced from the refinement results, were
99.9% purity in the desired proportion. The phase purity and the summarized in Table 1.
single phase nature of our compounds were confirmed by X-ray Since the Pr3þ and Gd3þ ions have the same valence, the Mn3þ/
powder diffraction (XRD) using (CuKa) radiation source. For elec- Mn4þ ratio is not affected. In Fig. 1-b, we have plotted the evolution
trical measurements, the samples, prepared as pellets, with a of V and 〈rA〉 as a function of the Gd doping rate. It is clear that the
thickness of 1 mm were used. In both side of each pellet, a thin substitution of Pr by Gd leads to an increase of the unit cell volume
silver layer was deposited in order to obtain ohmic contacts. An in contrary to what obtained in a previous work where Pr0.5-
Agilent 4294A analyzer was used to provide the conductance xGdxSr0.5MnO3 (x ¼ 0, 0.05 and 0.1) samples were prepared by
measurements over a wide range of frequency (40 Hze10 MHz). the conventional solid state route. The 〈rA〉 values decreases due to
The amplitude of the applied alternating signal is of 50 mV. A liquid the difference between the ionic radius of Pr ð rPr3þ ¼ 1:179 AÞ and
nitrogen cooled VPF-100 cryostat of Janis Corporation is used in Gd ð rGd3þ ¼ 1:107 AÞ [18]. In our samples prepared by sol-gel route,
order to provide the variation of temperature from 80 to 320 K. the increase of the Gd content inside the structure causes an in-
crease of the average 〈MneO〉 bond length and the decrease of the
3. Results and discussion average angle 〈MneOeMn〉 (listed in Table 1) and leads to a tilting
of the MnO6 octahedrons. As a result, the unit cell volume increases.
3.1. Structural properties
Fig. 1-a shows the XRD patterns for all our studied Pr0.5- 3.2. DC conductivity analysis
Table 1
The principle refinement results: the cell lattice parameters, the unit cell volume,
the average ionic radius in A-site 〈rA〉 and the cationic disorder in A-site s2 for Pr0.5-
xGdxSr0.5MnO3 (0 x 0.3) compounds.
sDC :T ¼ s0 :ekT
Ea
(1) sDC
sAC ðuÞ ¼ (2)
1 þ u2 t2s
where s0 is a constant, Ea is the activation energy and k the
Boltzmann constant. It can be seen that these plots are linear over a ts represents the relaxation time describing the electron-phonon
wide temperature range, confirming the thermally activated small scattering. From the fit shown in Fig. 3 of sAC(u) using relation
polaron hopping mechanism (SPH) [19]. The activation energies Ea (2), we have plotted the temperature evolution of ts for all our
are obtained from the slope of Ln(sDC.T) vs 1000/T curves. At high compounds. As shown in Fig. 5, the relaxation time increases with
temperatures, these values are 31, 35, 36 and 67 meV for x ¼ 0, 0.1, temperature which confirms the metallic behaviour of our samples
0.2 and 0.3 respectively. For x ¼ 0.3, the activation energy was in the studied range of temperature. This metallic character is
calculated in the temperature range of the insulator behavior. observed in other materials such as Li0.5Mn0.5Fe2O4-d [23] and Pr0.8-
The low value of activation energies confirms that the mobility
xBixSr0.2MnO3 [24].
of the charge carriers is due to a hopping mechanism dominated by In order to determine the AC activation energies, we have
the motion of electrons through the structure. The increase of Ea plotted Ln(ts) as a function of 1000/T relative to our compounds
values with Gd content confirms that the increase of electronic (the curves are not shown). The calculated values are 16, 17 and
localization makes the hoping of carriers from one site to another 18 meV for x ¼ 0, 0.1 and 0.2 respectively. These values are lower
harder. The introduced disorder by Gd substitution enhances than those obtained from DC measurements. Such difference is due
considerably the potential barrier that charge carrier has to over- to the fact that in the AC conductivity, the electronic jumps be-
come when x increase from 0.2 to 0.3 and this explains the strong tween localized states are frequency dependent [25]. An increase in
decrease of the conductivity. frequency is accompanied by an increase of the absorbed energy
[26] and this enhances the electronic jumps and reduces the acti-
3.3. AC conductivity analysis vation energy [27]. For x ¼ 0.3, we have a metallic behaviour at high
temperature and it is inappropriate to model the corresponding
The frequency dependence of the AC conductivity sAC for all our curve by a thermal activated process.
compounds at different temperatures is given in Fig. 3. This figure It is worth noticing that the semiconductor character observed
726 A. Ben Jazia Kharrat et al. / Journal of Alloys and Compounds 741 (2018) 723e733
Fig. 3. Frequency dependence of the conductivity at different temperatures for Pr0.5-xGdxSr0.5MnO3 (0 x 0.3) compounds: x ¼ 0.0 (a), x ¼ 0.1 (b), x ¼ 0.2 (c) and x ¼ 0.3 (d).
Table 2
The best fitting results of the conductivity sDC obtained from experimental data as a
function of frequency using the Jonscher's power law for Pr0.5-xGdxSr0.5MnO3
(0 x 0.3) compounds.
T (K) sDC A s R2 uh
(103 S cm1) (104 rd/s)
Pr0.3Gd0.2Sr0.5MnO3
80 4.06 5.68 1016 1.750 0.999 2214
90 6.53 3.68 1014 1.494 0.996 3385
95 8.42 1.15 1013 1.405 0.992 5405
Pr0.2Gd0.3Sr0.5MnO3
80 0.312 3.48 1014 1.445 0.998 772
90 0.725 3.22 1015 1.598 0.998 1271
95 0.976 1.56 1015 1.645 0.999 1483
100 1.67 1.20 1015 1.665 0.999 1965
105 2.49 8.23 1016 1.685 0.999 2553
110 3.56 2.49 1016 1.746 0.989 3424
115 4.9 2.26 1018 2.013 0.952 4155
6kT
s¼1 (5)
Wm s 1=s
Then, the plot of (1-s) vs temperature shown in Fig. 6-a, gives us uh ¼ DC
(8)
A
the average binding energy Wm of the PGSMO1 compound which is
evaluated as 22 meV. In Fig. 7, we have plotted the sAC data scaled by a Ghosh's model
Therefore, the most suitable model that describes the conduc- [35].
tion mechanism in the second sample is the non-overlapping small
polaron tunnelling model. According to this model, the s parameter
is given by Ref. [33]: sAC ðuÞ u
¼f (9)
sDC uh
4
s¼1þW (6) Accordingly, for the two studied compounds, the curves show a
kT
H
Lnðut0 Þ
perfect overlap for all temperatures indicating that the relaxation
In this expression, WH represents the polaron hopping energy. dynamics of the charge carriers is temperature independent [36].
For larger values of WH
, relation (6) can be reduced to: The variation of Ln(uh) with temperature for the doped samples
kT
is shown in the inset of Fig. 7. As we can see, these curves follow the
4 kT Arrhenius model and the calculated activation energies are 36 and
s¼1þ (7)
WH 43 meV for PGSMO1 and PGSMO2 respectively which suppose a
more localization of charge carriers due to the increase of Gd
s increases with increasing temperature indicating an increase in content.
728 A. Ben Jazia Kharrat et al. / Journal of Alloys and Compounds 741 (2018) 723e733
Fig. 7. Plot of (sAC/sDC) versus (u/uh) at different temperatures for PGSMO1 (a) and
PGSMO2 (b) compounds. The inset shows the temperature evolution of ln(uh) and the
calculated activation energies.
the resistance of bulk (inset of Fig. 8-a). The presence of a unique Nyquist plots, reported in Fig. 8, show a good agreement be-
semicircle in this structure can be attributed to bulk grain [41]. The tween experimental (scatter) and simulated curves (solid lines).
impedance of fractal capacitance CPEg is given by Ref. [42]: It is worth noticing that the Niquist plots for PGSMO1 and
PGSMO2 show a low-frequency inductive loop which increases
1 with temperature. This phenomenon was observed in metal/yttria
ZCPEg ¼ (10)
Q ðjuÞag stabilized zirconia electrodes and was attributed to the presence of
adsorbed species [43]. In our samples, the inductive effect can be
In this expression, Q is the capacitance value of the CPEg
attributed to the presence of gadolinium in the structure.
impedance and ag (0<ag < 1) describes the deviation from Debey's
For the PGSMO2 sample, and for 105 T 115 K, the inductive
model.
effect is more accentuated as seen in the second inset of Fig. 8-c.
Then, the real and imaginary parts of the total impedance of this
The change in the electronic model relative to this sample may
circuit are determined according to the following expressions:
originate from the increase of the L value with increasing
a p temperature.
0 Rg 1 þ Rg Q uag cos 2g
Z ¼Rþ a p2 a p2 (11) Fig. 9 indicates the frequency dependence of the real part Z0 of
1 þ Rg Q uag cos 2g þ Rg Q uag sin 2g the impedance at different temperatures referred to PGSMO1 and
PGSMO2 samples. In the low frequency range, Z0 shows higher
and values which decreases with temperature. For high frequency
a p values, Z0 presents lower values which may due to the release of
00 R2g Q uag sin 2g space charge polarization [44] and correlated with the increase of
Z ¼ a p2 a p2 Lu (12)
AC conductivity in this region. The Z0 spectra is well fitted by
1 þ Rg Q uag cos 2g þ Rg Q uag sin 2g
relation (11) for PGSMO1 and by relations (13) and (15) for
For the PGSMO2 sample, and in the temperature range PGSMO2.
80 T 100 K, The adequate equivalent circuit is illustrated in the The frequency dependence of the imaginary part Z00 of the
inset of Fig. 8-b where CPEgb is the fractal capacitance relative to impedance at various temperatures is shown in Fig. 10. These
grain boundaries. curves were simulated using relation (12) for sample PGSMO1 and
In this case, the real and imaginary parts of the total impedance equations (14) and (16) for sample PGSMO2 (solid lines). For each
are given by the expressions: compound, the plots show peaks which shift to higher frequencies
when the temperature increases showing a thermally activated
Rg relaxation phenomenon. A broad peak is observed in all Z00 which
Z0 ¼ R þ 2
1 þ Rg Cg u can be attributed to the existence of multiple relaxation times in
h i the two studied compounds. We can notice, in addition, that the
a p
Rgb 1 þ Rgb Q uagb cos gb2 maximum of these peaks decreases with increasing temperature
þ
a p 2 a p 2 and merges at high frequencies. Such behavior is a signature of an
1 þ Rgb Q uagb cos gb2 þ Rgb Q uagb sin gb2 accumulation of space charge polarization effect in our compounds
(13) which occurs at higher temperatures and lower frequencies [45]. By
analyzing both curves of Z0 and Z00 , we can notice that when the
and frequency increases, Z00 increases up to a particular value at which a
sudden decrease in Z0 value is detected. Such trend confirms the
00 R2g Cg u existence of a relaxation phenomenon in PGSMO1 and PGSMO2
Z ¼ 2 structures which occurs at a frequency noted fr. Using the relation
1 þ R g Cg u
t ¼ 1/(2pfr), the relaxation time t is calculated for the
R2gb Q uagb sin ap
2 Pr0.2Gd0.3Sr0.5MnO3 sample and plotted vs 1000/T in the inset of
þ 2 Lu Fig. 10. The obtained activation energy corresponding to non-Debey
a a p a p 2
1 þ Rgb Q u gb cos 2 gb
þ Rgb Q uagb sin gb2 relaxation is Ea ¼ 46 meV which is in good agreement with that
(14) obtained from Ln(uh) vs temperature curve.
The activation energies obtained from DC and AC measurements
For 105 T 115 K, The adequate equivalent circuit is shown in in the PGSMO1 and PGSMO2 compounds are different which sug-
the inset of Fig. 8-c. As a consequence, the real and imaginary parts gests that the charge carriers in the semiconductor region have to
of the total impedance of this circuit can be written as: overcome different energy barrier while conducting and relaxing.
For the Pr0.2Gd0.3Sr0.5MnO3 sample, the evolution of Rg and Rgb
Rg Rgb
Z0 ¼ 2 þ 2 (15) at different temperatures can be analysed using the adiabatic small
1 þ Rg Cg u 1 þ Rgb Cgb u polaron hopping model [46] described by:
Ea
and Rg;gb ¼ R0;1 T a exp (17)
kT
00 R2g Cg u R2gb Cgb u In this expression, Ea is the activation energy, R0 and R1 are
Z ¼ 2 þ 2 L u (16)
1 þ R g Cg u 1þ R C u
constants related to the polaron hopping concentration and
gb gb
The extracted parameters, for all these circuits are summarized Table 3
in Table 3 (for sample PGSMO1) and Table 4 (for sample PGSMO2). The equivalent circuit parameters obtained for the Pr0.3Gd0.2Sr0.5MnO3 sample.
It is worth noticing that Rgb present higher values due to accu-
T(K) L (106H) R(U) Rg(U) Q (1010F) ag
mulation of defects at the vicinity of grain boundaries such as
dangling bonds and a non-stoichiometric distribution of oxygen 80 2.00 30.3 130 3.63 0.942
90 1.94 37.6 61.5 7.27 0.984
[39].
730 A. Ben Jazia Kharrat et al. / Journal of Alloys and Compounds 741 (2018) 723e733
Table 4
The equivalent circuit parameters obtained for the Pr0.2Gd0.3Sr0.5MnO3 compound.
Fig. 10. Frequency dependence of the imaginary part (Z00 ) of the impedance at different
temperatures for (a) Pr0.3Gd0.2Sr0.5MnO3 and (b) Pr0.2Gd0.3Sr0.5MnO3 (The inset of this
figure shows the temperature dependence of the relaxation time). The solid line
Fig. 9. Frequency dependence of the real part (Z0 ) of the impedance at different represents the fit of the experimental data.
temperatures for: Pr0.3Gd0.2Sr0.5MnO3 (a) and Pr0.2Gd0.3Sr0.5MnO3 (b) compounds. The
solid line represents the fit of the experimental data.
Rgb
diffusion and a ¼ 1 in the case of adiabatic hopping. b¼ (18)
Rg þ Rgb
Fig. 11 shows the evolution of Ln(Rg/T) and Ln(Rgb/T) against
the inverse of absolute temperature. The calculated activation en- As depicted in Fig. 12, the blocking factor decreases with
ergies are 83 and 27 meV respectively. These values, compared to increasing temperature. Since b can be related to the fraction of
those obtained from DC conduction analysis and from Ln(t) vs charge carriers blocked at the grain boundaries, we can say that the
1000/T for PGSMO2 sample, shows that the effect of the grain charge carriers become freer as temperature increases and the total
boundaries dominates at low temperatures, whereas the effect of conductivity tends to the conductivity of grains.
the grains manifests towards the high temperatures. This behavior
indicates an increase of the mobility of charge carriers with rise in 3.5. Dielectric properties
temperature which causes an increase in the conduction process.
In order to study the grain boundaries effect on the total con- In the studied samples, the complex permittivity is defined as:
00 00
ductivity of Pr0.2Gd0.3Sr0.5MnO3 sample, we have calculated the ε* ðf Þ ¼ ε0 ðf Þ jε ðf Þ where ε0 and ε are respectively the real and the
blocking factor b expressed as [47]: imaginary parts of ε* . ε0 and ε00 are calculated using the following
relation [48]:
A. Ben Jazia Kharrat et al. / Journal of Alloys and Compounds 741 (2018) 723e733 731
e
ε* ¼ (19)
juε0 SZ *
S is the area of electrodes, e the thickness of the sample and
ε0 denotes the permittivity of the vacuum.
Fig. 13-a shows the frequency evolution of ε0 at some repre-
sentative temperatures for the PGSMO2 sample. As clearly seen,
very higher values of permittivity can be obtained with this type of
perovskites which can give higher values of capacities which is
beneficial for technological applications. The ε0 values increase with
temperature but decrease with frequency. It is well known that at
low frequencies and temperatures, all mechanisms of polarization
(interfacial polarization, dipoles, ions and electrons) can exist in the
structure. All polarization mechanisms tend to follow the applied
electric field. Consequently, a maximum of the real dielectric Fig. 13. Frequency dependence of: ε’ (a) and Ln (ε00 ) (b) at some representative tem-
permittivity is obtained at low frequencies and temperatures. As peratures for Pr0.2Gd0.3Sr0.5MnO3 compound. The inset shows the evolution of the
frequency increases, the polarized dipoles present in these struc- maximum barrier height versus temperature relative to the same compound.
tures cannot follow the orientation of the applied electric field and
the permittivity decreases strongly. An increase in temperature
conductivity and leads to the reduction of interfacial polarization.
causes an increase in the disorder of the dipoles and makes the
As a result, ε0 values decrease with frequency and temperature.
jump of charge carriers easier which causes an increase in the
In Fig. 13-b, we report the evolution of the complex permittivity
ε00 as a function of frequency at different temperatures for the
PGSMO2 sample. The plotted curves show straight lines whose
slopes are close to 1, which means that the DC conduction
mechanism is predominant in both doped compounds. The ε00 plots
can be analyzed with the equation given by Giuntini theory [49]:
00
4 m
ε ðuÞ ¼ ðε0 ε∞ Þ2p2 Nðne=ε0 Þ3 kT tm
0 WM u (20)
In this expression, N and n represent respectively the number of
the localized states and the number of charge carriers able to cross
the potential barrier, ε0 and ε∞ are the values of the static
(determined at low frequency) and high frequency dielectric con-
stant, t0 is the relaxation time and WM denotes the energy required
for carriers to jump over the potential barrier.
Relation (19) can be expressed as:
00
ε ðuÞ ¼ aðTÞ: um (21)
Fig. 12. Thermal evolution of the blocking factor b for Pr0.2Gd0.3Sr0.5MnO3 compound. where
732 A. Ben Jazia Kharrat et al. / Journal of Alloys and Compounds 741 (2018) 723e733
dielectric relaxation behavior of the hybrid polyvanadate (H3N(CH2)3NH3) organic semiconductor: Alizarin, Phys. B 388 (2007) 118e123.
[V4O10], Mater. Res. Bull. 48 (2013) 1978e1983. [46] A. Banerjee, S. Pal, E. Rozenberg, B.K. Chaudhuri, Adiabatic and non-adiabatic
[41] A.K. Behera, N.K. Mohanty, B. Behera, P. Nayak, Impedance properties of small-polaron hopping conduction in La1xPbxMnO3þd (0.0x0.5)-type ox-
0.7(BiFeO3)e0.3(PbTiO3) composite, Adv. Mater. Lett. 4 (2013) 141e145. ides above the metalesemiconductor transition, J. Phys. Condens. Matter 13
[42] A.K. Jonscher, The interpretation of non-ideal dielectric admittance and (2001) 9489.
impedance diagrams, Phys. Status Solidi (a) 32 (1975) 665e676. [47] M.J. Verkerk, B.J. Middelhuis, A.J. Burggraaf, Solid State Ionics 6 (1982)
[43] R.J. Aaberg, R. Tunold, M. Mogensen, R.W. Berg, R. Odegrd, Morphological 159e170.
changes at the interface of the nickel-yitria stabilized zirconia point electrode, [48] A. von Hippel, Dielectrics and Waves, John Wiley and Sons, New York, 1954.
J. Electrochem. Soc. 145 (1998) 2244e2252. [49] J.C. Giuntini, J.V. Zanchetta, D. Jullien, R. Eholie, P. Houenou, Temperature
[44] M. Ganguli, M. Harish Bhat, K.J. Rao, Lithium ion transport in Li2SO4-Li2O-B2O3 dependence of dielectric losses in chalcogenide glasses, J. Non-Cryst. Solids 45
glasses, Phys. Chem. Glasses 40 (1999) 297e304. (1981) 57e62.
[45] K.P. Chandra, K. Prasad, R.N. Gupta, Impedance spectroscopy study of an